A/B test across full stack applications

A/B testing is a powerful method to test hypotheses and bring your best ideas to customers.

Combining web, mobile, and server-side feature flags with your customer data, Split makes it easy to experiment with any part of your application and maximize the impact of every feature. 

Fundamentals of A/B testing

A/B testing is an experimentation process where two variations of software are tested among a randomized, controlled set of users to ensure statistical rigor.

Each A/B test includes a set of metrics as part of the hypothesis, defining the goals of the experiment. In many cases you may want to affect several metrics while not affecting others, called guardrail metrics.

Understanding Experimentation Platforms book by O'Reilly.

Guide to full stack A/B testing

Turn every feature into an A/B test

Turning every feature you develop into an A/B test. Split ties feature flags to customizable metrics from your operational and customer data pipelines, automatically measuring the impact of every feature on every metric you define.

Split takes a statistically rigorous approach to A/B testing, detecting variation across treatment groups to determine whether an experiment was successful and if you should continue rolling it out.

Scale A/B testing with an experimentation platform

With a feature flag foundation for frontend, backend, and mobile apps connected to your existing data pipeline, Split powers A/B testing anywhere.